Squares from blocks of consecutive integers: a problem of Erdős and Graham (Q663569): Difference between revisions

Let \(n, k\in\mathbb{N}\) and put \[ f(n,k)=n(n+1)\cdots(n+k-1). \] An old result of \textit{P. Erdős} and \textit{J. Selfridge} [Ill. J. Math. 19, 292--301 (1975; Zbl 0295.10017)] states that the \(f(n,k)\) is never a perfect power of an integer, provided \(n\geq 1\) and \(k\geq 2.\) \textit{P. Erdős} and \textit{R. Graham} [Old and new problems and results in combinatorial number theory. Gene\`{v}e: L'Enseignement Mathématique, Université de Gene\`{v}e (1980; Zbl 0434.10001)] asked whether a more general Diophantine equation \[ \prod_{j=1}^{r}f(n_{j},k_{j})=x^2 \tag{1} \] could have, for fixed \(r\geq 1\) and \(\{k_{1},k_{2},\ldots,k_{r}\}\) with \(k_{j}\geq 4\) for \(j=1,2,\ldots,r\), at most finitely many solutions in positive integers \((n_{1},n_{2},\ldots,n_{r},x)\) satisfying the condition \[ n_{j}+k_{j}\leq n_{j+1}\tag{2} \] for each \(j\in\{1,\ldots,r-1\}\). The last condition guarantees that the corresponding blocks of integers are disjoint. The authors concentrate on the case \(k_{1}=k_{2}=\ldots=k_{r}=5\) and prove that if \(r\geq 5\) then there are infinitely many integers solutions of equation (1) satisfying (2), and thus answering the question of Erdős and Graham in the negative.